Magnetic graphene oxide-anchored Ni/Cu nanoparticles with a Cu-rich surface for transfer hydrogenation of nitroaromatics

Hongbin Shi,Qing Liu,Xiaofeng Dai,Teng Zhang,Yuling Shi,Tao Wang*

State Key Lab of Chemical Engineering,Department of Chemical Engineering,Tsinghua University,Beijing 100084,China

Keywords:Nitroaromatics Transfer hydrogenation Copper-nickel bimetals Core-shell nanoparticles Magnetic catalysts

ABSTRACT The bimetallic nanoparticles compositing of Ni-rich core and Cu-rich shell(Ni/Cu NPs)were successfully synthesized by a liquid-phase thermal decomposition method.The content of copper and nickel in Ni/Cu NPs was controllable by adjusting the ratio of two metal precursors,copper formate(Cuf)and nickel acetate tetrahydrate(Ni(OAc)2·4H2O).Ni/Cu NPs were further anchored on graphene oxide(GO)to prepare a magnetic composite catalyst,called Ni/Cu-GO.The dispersibility of Ni/Cu NPs in solution was enhanced by GO anchoring to prevent the sintering and aggregation during the reaction process,thereby ensuring the catalytic and cycling performance of the catalyst.The catalytic transfer hydrogenation(CTH)reaction of nitroaromatics was investigated when ammonia borane was used as the hydrogen source.Cu dominated the main catalytic role in the reaction,while Ni played a synergistic role of catalysis and providing magnetic properties for separation.The Ni7/Cu3-GO catalyst exhibited the best catalytic performance with the conversion and yield of 99% and 96%,respectively,when 2-methyl-5-nitrophenol was used as the substrate.The Ni7/Cu3-GO catalyst also exhibited excellent cyclic catalytic performance with the 5-amino-2-methylphenol yield of above 90% after six cycles.In addition,the Ni7/Cu3-GO catalyst could be quickly recycled by magnetic separation.Moreover,the Ni7/Cu3-GO catalyst showed good catalytic performance for halogen-containing nitroaromatics without dehalogenation.

1.Introduction

Catalytic transfer hydrogenation (CTH) of nitroaromatics is an important way to synthesize aromatic amines.Compared with traditional direct hydrogenation process,the CTH is a more convenient and safer method due to the absence of dangerous highpressure hydrogen gas [1-3].The reaction could be carried out in the liquid phase in mild condition without pressurizing,therefore no expensive high pressure equipment was required.The diversity of hydrogen sources for CTH provides a new way to improve the selectivity of the reaction.Therefore,the CTH is a promising route for the hydrogenation of nitroaromatics.

The commonly used hydrogen sources are mainly alcohols[4,5],formic acid and its salts [6],hydrazine hydrate [7],ammonia borane [8],sodium borohydride [9],etc.Strong bases are often used as the co-catalysts when low reducing hydrogen sources such as alcohols or formic acid are used.The presence of strong bases is not suitable for base-sensitive substrates,such as halogencontaining substituents(such as F,Cl,Br,etc.),resulting in dehalogenation of the substrate [10,11].Moreover,the steric hindrance and electron-withdrawing groups (such as halogen substituents)in nitroaromatics containing multiple substituents make them difficult to be hydrogenated efficiently.Therefore,it is usually necessary to use a strongly reductive substance as a hydrogen source for nitroaromatics that are difficult to be reduced.Ammonia borane(NH3·BH3),a potential hydrogen source with hydrogen content as high as 19.6%（mass）,can exist stably in aqueous solution,making it an ideal high reductive hydrogen source [12,13].

Currently,the heterogeneous catalysts reported for the CTH reaction of nitroaromatics are mainly divided into two categories:noble metal catalysts such as Pd [14,15],Rh [16],Pt [17],Ru [18],etc.,and non-noble metal catalysts such as Ni [19,20],Cu [21,22],Fe[7],Co[23],etc.The advantages of noble metal catalysts are high catalytic activity and good stability,but the disadvantage of high cost is also obvious.Non-precious metal catalysts are potential alternatives of noble metal catalysts due to their wide availability and low cost,despite the disadvantages of low catalytic activity and poor cycle stability.Compared with single-metal catalysts,bimetallic catalysts have received extensive attention owing to the synergistic enhancement of different elements and tunable physicochemical properties.The synergistic enhancement is mainly reflected in the changes of geometric and electronic effects of the catalyst,resulting in more suitable physical and chemical properties than single-metal catalysts [24,25].Non-precious bimetallic catalysts are low-cost,efficient and stable,and have gradually become a research hotspot,such as Cu-Ni [26],Ni-Co[27],Fe-Ni[28],Fe-Cu[29],Co-Cu[30],etc.Copper-nickel bimetallic nanoparticles (Cu-Ni NPs) are considered as a promising catalyst with great potential and application prospect in the field of CTH for the following reasons.First,Cu-Ni NPs have shown obvious synergistic enhancement effect in the CTH reaction [31,32].Second,Cu-based catalysts have be proved to exhibit good cycle stability,because Cu still shows good CTH activity even if it is oxidized to Cu2O and CuO[33].In contrast,Co,Ni and Fe are almost deactivated after oxidation.Third,magnetic Cu-Ni NPs can be rapidly magnetically separated and recycled under an external magnetic field,which greatly facilitates their industrial application.Ni-Cu core-shell(Ni/Cu)NPs,a composite catalyst combining the strong magnetism of core Ni and the stable catalytic properties of shell Cu,are potentially ideal magnetic non-noble bimetallic catalysts.However,there are few studies on the preparation and application of Ni/Cu NPs.

The dispersion of metal nanoparticles in the reaction solution has a great influence on the catalytic performance.In our previous study [34],the treatment of CuNi alloy NPs with NOBF4greatly improved the dispersibility of nanoparticles in the reaction solution,resulting in a good catalytic performance.However,unsupported nanoparticles are easily chemically sintered and aggregated by strongly reductive substance such as ammonia borane,thereby degrading the catalytic performance.Therefore,metal nanoparticles are often loaded onto supports to improve the dispersion and increase the exposure of active sites [35].Graphene oxide (GO) is a common carrier for metal nanoparticles loading[36-38],and metal nanoparticles are firmly anchored on the GO through surface oxygen bonding,preventing the chemical sintering and agglomeration of nanoparticles during the CTH reaction.In addition,the adsorption capacity of the graphene oxide surface is very strong,which can enrich the reactants near the catalyst and further improve the catalytic performance.Compared with porous carrier,GO with the planar structure make the supported nanoparticles catalysts have less diffusion resistance.

In this study,the nanoparticles compositing of Ni-rich core and Cu-rich shell (Ni/Cu NPs) with different compositions were successfully prepared by liquid-phase thermal decomposition using copper formate (Cuf) and nickel acetate tetrahydrate (Ni(OAc)2·4-H2O) as metal precursors.The Ni/Cu NPs were further anchored on GO to synthesize Ni/Cu-GO composite catalyst.During the synthesis of Ni/Cu NPs,oleylamine (OAM) was used as a complexing agent and stabilizer,and liquid paraffin was used as solvent.2-methyl-5-nitrophenol was used as a template substrate to investigate the CTH activity of Ni/Cu-GO catalyst with ammonia borane as the hydrogen source.Ni7/Cu3-GO catalyst exhibited the best performance.The recycle and reuse of the catalyst could be realized by magnetic separation.The Ni7/Cu3-GO catalyst showed the good recycling performance owing to the good dispersion and stable core-shell structure of Ni/Cu NPs,which was benefited from GO anchoring.Furthermore,the prepared magnetic Ni7/Cu3-GO catalyst also displayed the high activity for transfer hydrogenation of various nitroaromatics.

2.Experimental

2.1.Chemical and materials

Copper formate tetrahydrate (Cuf·4H2O,AR) and nickel acetate tetrahydrate (Ni(OAc)2·4H2O,AR) were purchased from Alfa Aesar(China) Chemical Co.,Ltd.and Shanghai Macklin Biochemical Co.,Ltd.,respectively.Oleylamine (OAM,90%) was purchased from Shanghai Aladdin Biochemical Technology Co.,Ltd.Liquid paraffin(AR) was purchased from Sinopharm Chemical Reagent Co.,Ltd.Hexane (AR),anhydrous ethanol (AR),and anhydrous methanol(MeOH,AR) were purchased from Shanghai Titan Technology Co.,Ltd.Ammonia borane (NH3·BH3,97%) was purchased from Shanghai Acmec Biochemical Co.,Ltd.2-methyl-5-nitrophenol(98%)was purchased from Meryer (Shanghai) Chemical Technology Co.,Ltd.Copper oxide (CuO,AR) was purchased from Beijing Modern East Technology Development Co.,Ltd.,and nickel oxide (NiO,99.5%,diameter 25-30 μm) was purchased from Bide Pharmatech Ltd.Large-diameter graphene oxide sheets (GO,sheet diameter ＞5 μm,number of layers 1-6) were purchased from Jiangsu XFNANO Materials Technology Co.,Ltd.Copper formate tetrahydrate was dried in a vacuum drying oven at 90 °C for 12 h to obtain copper formate (Cuf).

2.2.Characterization

The morphologies of Ni/Cu NPs and Ni/Cu-GO as well as the dispersion of Ni/Cu NPs on GO were observed by transmission electron microscopy (TEM,HT-7700,Hitachi).The crystal structure of the catalysts was characterized by X-ray diffraction (XRD,D8 Advance,Bruker).The lattice fringes and elemental distribution of nanoparticles were analyzed by high-magnification transmission microscopy (HRTEM,JEM2100F,JEOL) coupled with energy dispersive spectrometer (EDS).Element valence on the surface of nanoparticles was characterized by X-ray photoelectron spectroscopy (XPS,ESCALAB 250Xi,Thermo Fisher).Raman spectra analysis of GO and Ni/Cu-GO catalysts was performed using a high-resolution Raman spectrometer (Raman,LabRAM HR Evolution,HORIBA).The product composition and concentration were analyzed by high performance liquid chromatography (HPLC,1260 Infinity,Agilent) under the following conditions.Column:reversed-phase column (Eclipse Plus C18,5 μm,4.6 mm × 250 m m);detector: UV detector (I=230 nm);detection temperature:30 °C;mobile phase: acetonitrile aqueous solution,gradient elution,0-10 min,acetonitrile volume content increased from 40%to 55%,10-15 min,acetonitrile volume content decreased from 55%to 40%;flow rate:1.0 ml·min-1.Mass spectrometry data were obtained by HPLC-MS data system (6460 QQQ,Agilent).Aromatic amines and other compounds in the product were identified by liquid chromatography retention index and mass spectrometry in comparison to authentic standards,literature and library data.

2.3.Ni/Cu NPs synthesis

Ni/Cu NPs were synthesized by a sequential reduction method.The details was illustrated by taking the synthesis of Ni5/Cu5NPs as an example.In the first step,2 mmol of Ni(OAc)2·4H2O,2.4 g of OAM (OAM/metal molar ratio of 4:1) and 15 ml of liquid paraffin were added to a 50 ml three-necked flask,heating at 80 °C under stirring until all dissolved to give a homogeneous solution.The solution was then gradually heated to 220 °C and reacted at 220 °C for 30 min.After the reaction,the mixture was cooled to room temperature by removing the heat source.In the second step,2 mmol Cuf was dissolved in 2.4 g OAM and 5 ml liquid paraffin at 60 °C,and was added to the reaction solution obtained from the first step.The mixture was mechanically stirred at 300 r·min-1until uniform,then gradually heated to 220 °C and held at 220 °C for 30 min.A nitrogen atmosphere was maintained throughout the reaction.Subsequently,the reaction solution was cooled to room temperature and magnetically separated.The solid products were washed twice with a 1:1 mixture of hexane/ethanol and then dispersed in hexane for further use.The Ni/Cu NPs with different compositions were synthesized by changing the molar ratio of Cuf to Ni(OAc)2·4H2O while the total molar amount of metal precursors was kept as 4 mmol.

2.4.Ni/Cu-GO catalyst synthesis

In a typical synthesis process of Ni/Cu-GO catalyst,10 ml of Ni/Cu NPs in hexane (2 mg·ml-1) with was added dropwise to 80 ml of GO in ethanol(0.5 mg·ml-1).The mixture was ultrasonically dispersed for 1 h to ensure that the NPs were fully loaded on the GO[39,40].The mixture was then magnetically separated,washed twice with ethanol,and stored in ethanol for further use.

2.5.CTH reaction of nitroaromatics catalyzed by Ni/Cu-GO

In a typical CTH reaction,nitroaromatics (0.5 mmol),Ni/Cu-GO catalyst(30%(mol),catalyst amount was defined as mole of metal per mole of reactant),3 ml methanol and 7 ml of water was added to the round bottom flask and mixed for 5 min at room temperature.The flask was sealed with a balloon after addition of ammonia borane (1.5 mmol).The mixture was stirred with a shaker at 200 r·min-1at room temperature and reacted for 30 min.After the CTH reaction,the Ni/Cu-GO catalyst was separated by a strong magnet for about 5 min,washed twice with ethanol and separated again magnetically,and then it could be put into the next catalytic reaction.The separated supernatant was diluted with methanol as crude product and directly analyzed by HPLC-MS.

When he came to the outskirts26 of the wood he said to his followers27: You wait here, I ll manage the giants by myself ; and he went on into the wood, casting his sharp little eyes right and left about him

3.Results and Discussion

3.1.Structure and composition of Ni/Cu NPs

As shown in Fig.1,Ni/Cu NPs were synthesized by sequential reduction method.First,the Ni(OAc)2·4H2O-OAM complex decomposed at 220 °C to produce Ni NPs.Then,the Cuf-OAM complex decomposed during the heating process in the second step,accompanied by the galvanic displacement reaction of the Cuf-OAM complex with the Ni NPs,resulting in the formation of Cu shell on the surface of the Ni NPs.Subsequently,the thermal decomposition of Nif-OAM complex generated by the galvanic displacement reaction lead to Ni doping of the Cu shell,meanwhile the outer copper and inner nickel partially dissolved into each other,so as to form the alloyed shell of nanoparticles.

In order to confirm the core-shell structure of the nanoparticles,the elemental distribution of the nanoparticles was analyzed by STEM-EDS elemental mapping-scan (Fig.2(a)) and line-scan (Fig.2(b)) by taking Ni5/Cu5NPs as an example.An obvious core-shell structure could be observed from Fig.2(a),Ni element was mainly distributed in the core of the particle while Cu element was deposited on the outer circle of Ni core.It was worth noting that there was still partial Ni doping in the Cu shell,indicating that the shell was copper-rich alloy,as shown in Fig.2(b).

Fig.3 exhibited the TEM images and particle size distributions of Ni/Cu NPs with different Cu/Ni ratios.Ni9/Cu1,Ni7/Cu3,Ni6/Cu4,Ni5/Cu5,Ni4/Cu6and Ni3/Cu7NPs were obtained by changing the molar ratio of Ni(OAc)2·4H2O/Cuf.The copper and nickel content of the Ni/Cu NPs was basically consistent with the ratio of precursors by ICP analysis (Table 1).The particle size of Ni/Cu NPs with different Cu/Ni ratios ranged from 40 to 45 nm,and the average diameter and uniformity of particle size did not change significantly.It was worth noting that the surface of the obtained nanoparticles became uneven when the Cu/Ni ratio was greater than 5:5(Fig.3(e),(f)).It was probably because the decomposition rate of a large amount of Cuf-OAM was so fast that the deposition of Cu on the Ni NPs was not uniform.When the ratio of Cu/Ni was 7:3 (Fig.3(f)),there were even a small amount of Cu NPs with an average particle size of about 7.8 nm in the product,which were mainly pure Cu NPs generated by thermal decomposition of Cuf-OAM.

Table 1 ICP analysis of Ni/Cu NPs with different Cu/Ni ratios

Fig.4 showed the XRD patterns of pure Cu,Ni3/Cu7,Ni5/Cu5,Ni7/Cu3and pure Ni NPs.Comparing the XRD patterns with the standard card,the crystal structure of Ni/Cu NPs were close to the face-centered cubic (fcc) crystal structure of the inner core Ni.The diffraction peaks around 44.6°,52.0° and 76.6° corresponded with the (111),(200) and (220) crystal planes of fcc Ni,respectively.With the increase of the Cu/Ni ratio in the NPs,the position of each peak also shifted to the Cu peak accordingly.In the XRD pattern of Ni3/Cu7NPs,the diffraction peaks near 43.3°,50.4° and 74.1° were consistent with the (111),(200) and (220) crystal planes of fcc Cu,intimating that the surface Cu layer reached a certain thickness.Pure Cu NPs were easily oxidized,so a diffraction peak representing the(111)crystal plane of Cu2O appeared around 37.1°[41].However,the Cu2O peak significantly weakened or even disappeared in the XRD patterns of Ni/Cu NPs,indicating that Ni/Cu NPs were not easily oxidized.

3.2.Structure of Ni/Cu-GO

Fig.1.Schematic illustration of the synthesis mechanism of Ni/Cu NPs.

Fig.2.(a) HAADF-STEM image and STEM-EDS elemental mapping-scan image,and (b) HAADF-STEM image and STEM-EDS elemental line-scan image of Ni5/Cu5 NPs.

Fig.3.TEM images and particle size distributions of Ni/Cu NPs with different Cu/Ni ratios: (a) Ni9/Cu1;(b) Ni7/Cu3;(c) Ni6/Cu4;(d) Ni5/Cu5;(e) Ni4/Cu6;(f) Ni3/Cu7.

The prepared Ni/Cu NPs with different Cu/Ni ratios were anchored on GO to form composite catalysts for the purpose of keeping the NPs well dispersed during the reaction.Taking the Ni7/Cu3-GO as an example for characterization,the HAADF-STEM image and the STEM-EDS elemental mapping-scan image demonstrated the remarkable Ni-Cu core-shell structure,as shown in Fig.5(a).Fig.5(b) showed the TEM image of the Ni7/Cu3-GO,exhibiting that the Ni7/Cu3NPs were well dispersed on GO without obvious aggregation.The Ni7/Cu3NPs was anchored on GO mainly because of the unique two-dimensional planar structure and high specific surface area of GO.Moreover,the abundant polar oxygen-containing functional groups such as epoxy groups,carboxyl groups,and hydroxyl groups on the surface of GO providedsites for van der Waals forces and coordination interactions with metal nanoparticles [42-44].

Fig.4.XRD patterns of Ni/Cu NPs with different Cu/Ni ratios.

Fig.6(a) and Table 2 exhibited the Raman data of GO and Ni7/Cu3-GO,both showing obvious graphene oxide spectral shapes.It could be seen that the D peak at about 1350 cm-1represented the Raman peak of disordered features of GO and Ni7/Cu3-GO due to structural defects,while the G peak at about 1590 cm-1represented the first-order scattering owing toE2gvibrational modes inside the carbon mesh [45,46].It was found that the ratio ofID/IGwas slightly enhanced after loading Ni7/Cu3NPs by comparing the Raman spectra of Ni7/Cu3-GO with GO,which proved that the defects of substrate graphene oxide had increased after nanoparticles loading [47,48].Fig.6(b) showed the XRD patterns of Ni7/Cu3,GO and Ni7/Cu3-GO.The peaks of Ni7/Cu3-GO matched those of Ni7/Cu3NPs and GO except a slight shift in the position of the main peak of GO,indicating that the catalyst was a composite of two components.The shift of the GO peak could be attributed to the metal-anchored doping on the surface of the GO after nanoparticle loading,resulting in an increase in the interlayer distance of GO,which was reflected in the left shift of the XRD diffraction peak[47,49].

3.3.Catalytic performance of Ni/Cu-GO

Table 3 exhibited the catalytic performance of Ni/Cu-GO with different Cu/Ni ratios in the CTH reaction of 2-methyl-5-nitrophenol when ammonia borane was used as the hydrogen source.Cu provided the main catalytic activity,since both the prepared Cu-GO catalyst and the commercial CuO powder showed a certain catalytic activity with the yields of 5-amino-2-methylphenol of 78% and 60%,respectively (Table 3,entries 4-5).However,the commercial NiO powder showed almost no catalytic activity,and the prepared Ni-GO catalyst was deactivated due to its easy oxidation in aqueous systems (Table 3,entries 6-7).

Table 2 Raman spectra of GO and Ni7/Cu3-GO

Table 3 Transfer hydrogenation performance of various Cu-Ni composite catalysts

Fig.5.(a) HAADF-STEM image and STEM-EDS elemental mapping-scan image of Ni7/Cu3 NPs.(b) TEM image of Ni7/Cu3-GO catalyst (the inner image was the STEM-EDS elemental line-scan image of a Ni7/Cu3 NP).

Fig.6.(a) Raman spectra of GO and Ni7/Cu3-GO;(b) XRD patterns of Ni7/Cu3,GO and Ni7/Cu3-GO.

3.4.CTH of nitroaromatics by Ni7/Cu3-GO

In the CTH with ammonia borane as the hydrogen source,Cu played a leading role in catalysis,and Ni took a synergistic enhancement effect and provided the required magnetic properties for magnetic separation.In practical reactions,both metallic Cu and oxidized Cu could provide catalytic active sites for the reaction.The oxidized Cu species on the surface of nanoparticles would transform into reduced Cu at the initial stage of the reaction to play a catalytic role[47,52,53].The CTH of nitroaromatics was based on the Langmuir-Hinshelwood mechanism [31],where the catalyst acted as an adsorption site for nitroaromatics and ammonia borane.The hydrogenation of nitro groups in aromatics was achieved by the transfer of electron and hydrogen species.So far,there was no unified conclusion on the hydrogen transfer mechanism when ammonia borane was used as the hydrogen source,because the reaction path could be easily changed with the reaction conditions[13].In alcohol/water solution systems,the general mechanistic explanation was that the negative hydrogen of borane in ammonia borane and the proton in polar solvent molecules provided dihydrogen atoms in a hydrogen transfer process,respectively.The borate NH4B(OH)4was a by-product of the CTH reaction [54,55].Therefore,the catalytic performance was affected by the solvent,the catalyst dosage and the ammonia borane dosage.

Fig.7.The Cu 2p and Ni 2p XPS spectra of Ni7/Cu3,Ni5/Cu5 and Ni3/Cu7 NPs.

Table 4 (entries 1-7) showed the effect of solvent on the CTH reaction.On the one hand,the solvent affected the adsorption and desorption of substrates and products on the catalyst surface[56].On the other hand,the dehydrogenation capacity of the protic solvent affected the rate of transfer hydrogenation when ammonia borane was the hydrogen source [13,54].The dehydrogenation of water was easier than that of alcohol,but a certain amount of organic solvent was also needed to facilitate the dissolution of the substrate.When using 30%(vol) methanol aqueous solutionas solvent,the yield of 5-amino-2-methylphenol reached 94% in 10 min (Table 4,entry 5).In contrast,when pure methanol was used as solvent,the conversion and yield were only 11% and 8%even if the reaction times was extended to 30 min (Table 4,entry 1).

Table 4 Ni7/Cu3-GO-catalyzed CTH of 5-nitro-2-methylphenol under different conditions

Fig.8.(a) Catalytic performance of Ni7/Cu3-GO in 6 cycles.(b) Comparison of catalytic performance of Ni7/Cu3-NOBF4 and Ni7/Cu3-GO in 6 cycles.(c) Magnetization versus applied magnetic field for Ni7/Cu3-GO at 25°C.(d)Magnetic separation and redispersion of Ni7/Cu3-GO in ethanol.Reaction conditions:2-methyl-5-nitrophenol(0.5 mmol),NH3·BH3 (1.5 mmol),MeOH: H2O=3:7 (volume ratio,10 ml),catalyst/substrate (30% (mol)),25 °C,30 min.

Fig.9.TEM images:(a)Ni7/Cu3-NOBF4 before reaction.(b)Ni7/Cu3-NOBF4 after 3 cycles.(c)Ni7/Cu3-GO before reaction.(d)Ni7/Cu3-GO after 5 cycles(the inner image was the STEM-EDS elemental line-scan image of a Ni7/Cu3 NP).

Table 4(entries 5 and 8-11)also exhibited the effect of catalyst dosage on the CTH reaction,where the catalyst dosage was calculated as the metal/substrate molar ratio.When the catalyst dosage was 10% (mol) and 20% (mol),the reaction proceeded slowly,giving the yields of 5-amino-2-methylphenol were 36% and 64% at 30 min,respectively (Table 4,entries 8-9).However,the reaction did not reach final equilibrium at 30 min.When the catalyst dosage was increased to more than 30%(mol),the reaction rate was faster and the final equilibrium state was reached earlier,resulting in the product yield of 94% at 10 min (Table 4,entries 5 and 10-11).

Table 4 (entries 12-16) showed the effect of ammonia borane dosage on the CTH reaction.The catalytic reaction did not occur without ammonia borane(Table 4,entry 12).When the molar ratio of NH3·BH3/substrate was 2,the yield of 5-amino-2-methylphenol was 68%(Table 4,entry 13),because ammonia borane was not sufficient to completely convert the substrate to the final amine product.When the molar ratio of ammonia borane/substrate was greater than 3,the yield of amine products reached ＞96%(Table 4,entries 14-16).

Table 5 Ni7/Cu3-GO-catalyzed CTH of other nitroaromatics

Table 6 Hydrogenation of nitroaromatics to aromatic amines using various catalysts

GO played a great role in maintaining high catalytic activity of Ni7/Cu3NPs in cycling reactions.The catalytic performance was compared for the Ni7/Cu3-GO and the unsupported Ni7/Cu3NPs treated with NOBF4(Ni7/Cu3-NOBF4).The cyclic performance of two catalysts was very distinct,as shown in Fig.8(a).The Ni7/Cu3-NOBF4exhibited good dispersion before the first reaction cycle(Fig.9(a)),and the yield of 5-amino-2-methylphenol in the initial reaction was 90%.However,the catalytic performance was significantly reduced as recycling,exhibiting the yield of 28% after three cycles.Moreover,the Ni7/Cu3-NOBF4catalyst suffered severe sintering and aggregation during reaction(Fig.9(b)).This was mainly because the metal ions on the surface of Ni7/Cu3-NOBF4,especially Cu ions,were probably reduced to metal in the presence of ammonia borane[57,58].As a result,the sintering and aggregation of Ni7/Cu3-NOBF4lead to a significant decrease in the catalytic performance.

In contrast,the Ni7/Cu3-GO exhibited the good cyclic performance,as shown in Fig.8(b).After six cycles,the catalytic activity of Ni7/Cu3-GO did not decrease significantly with the substrate conversion and amine yield of 99% and ＞90%,respectively.Fig.9(c),d showed the TEM images of the Ni7/Cu3-GO before reaction and after five cycles (Ni7/Cu3-GO-5).From the EDS element linescan in Fig.9(d),it could be seen that the NPs in the Ni7/Cu3-GO-5 still maintained a relatively stable structure with a Cu-rich shell and a Ni-rich core after five cycles of reaction.In addition,Ni7/Cu3NPs were still stably anchored on GO without sintering and aggregation,as indicated in Fig.9(d).The GO support not only played the role of anchoring the Ni7/Cu3NPs to prevent nanoparticle from sintering,but also promoted the reaction by adsorbing the substrate through π-π bonding to increase the substrate concentration around the catalyst[8,26,59].Furthermore,the Ni7/Cu3-GO catalyst could be magnetically separated due to the saturation magnetization of 11.3 emu/g at 25°C,as shown in Fig.8(c),(d).Although the Ni7/Cu3-GO was not superparamagnetic with a coercivity of 60 Oe,the steric hindrance formed by the large-sized GO sheets enabled the nanoparticles to be well redispersed after magnetic separation.Therefore,the cyclic catalytic performance of the Ni7/Cu3-GO did not decrease,unlike Ni7/Cu3-NOBF4.

Fig.10.XPS spectra of Ni7/Cu3-GO: (a) Cu 2p,(c) Cu LMM,(e) Ni 2p.XPS spectra of Ni7/Cu3-GO-5: (b) Cu 2p,(d) Cu LMM,(f) Ni 2p.

Fig.10(a),c showed the Cu 2p and Cu LMM XPS spectra of Ni7/Cu3-GO.In the Cu 2p spectrum,the binding energy peaks at 932.2 and 952.0 eV corresponded to Cu 2p3/2and Cu 2p1/2of Cu0/Cu+,while the main peaks at 933.5 and 954.1 eV with the satellite peaks at 942.3 and 962.1 eV were respectively assigned to Cu2+2p3/2and Cu2+2p1/2[31,34].In the Cu LMM spectrum,the peak at 570.0 eV representing Cu2O further confirmed the existence of Cu+,and the peak at 568.2 eV represented Cu and CuO [60].Combining the Cu 2p and Cu LMM spectra,it could be concluded that there were three species of Cu,Cu2O and CuO on the surface of the Ni7/Cu3-GO.The existence of Cu2O and CuO was mainly due to the partial oxidation of Cu during the sample preparation.

Fig.10(b),(d) displayed the Cu 2p and Cu LMM XPS spectra of Ni7/Cu3-GO-5.Compared with the fresh Ni7/Cu3-GO,the Cu species on the surface of the Ni7/Cu3-GO-5 changed significantly.The binding energy peaks at 932.9 and 952.8 eV represented 2p3/2and 2p1/2of CuO,while the main peaks at 934.6 and 955.0 eV with the satellite peaks at 943.3 and 963.0 eV corresponded to Cu(OH)22p3/2and Cu(OH)22p1/2,respectively.The disappearance of the peaks corresponding to Cu and CuO at the binding energy of 568.2 eV in the Cu LMM spectrum donated the absence of Cu and CuO on the surface of the Ni7/Cu3-GO-5.In previous studies [60],the binding energypeak of Cu(OH)2in the Cu LMM spectrum was 570.4 eV,which was very close to the binding energy of 570 eV for Cu2O.In summary,the Cu species on the surface of Ni7/Cu3-GO-5 mainly existed in the form of Cu2O and Cu(OH)2on account of the gradually oxidation of Cu.

Fig.10(e),f exhibited the Ni 2p XPS spectra of Ni7/Cu3-GO and Ni7/Cu3-GO-5.The binding energy peaks at 852.7 and 870.2 eV represented Ni02p3/2and Ni02p1/2,while the peaks at 855.2 and 873.1 eV with two satellite peaks at 859.8 and 878.9 eV corresponded to Ni2+2p3/2and Ni2+2p1/2[61].Therefore,the Ni on the surface of Ni7/Cu3-GO catalyst was almost completely oxidized to inactive Ni2+after five cycles of reaction.The change of Ni status would have a certain impact on the structure of the active site,affecting the synergistic effect between copper and nickel,which was manifested as a decrease in the catalytic performance.However,the Cu2O species on the surface of Ni7/Cu3-GO-5 provided the main catalytic activity for the subsequent reactions.

The Ni7/Cu3-GO exhibited good catalytic performance for other nitroaromatics,such as nitrobenzene,2-nitrotoluene,3-bromonitrobenzene,3-chloro-4-fluoronitrobenzene and 4-nitrophenol (Table 5,entries 1-5).In particular,halogencontaining nitroaromatics were very easy to be dehalogenated during hydrogenation,which not only greatly decreased the selectivity of amine product,but also caused serious corrosion to equipment.Unlike Pd-based catalysts which often lead to dehalogenation[62],Ni7/Cu3-GO exhibited the high selectivity for hydrogenation of nitro group without dehalogenation.For the transfer hydrogenation of 3-chloro-4-fluoro-nitrobenzene,the 3-chloro-4-fluoroaniline yield was high as 96% with Ni7/Cu3-GO as the catalyst(Table 5,entry 4).

Using 4-nitrophenol as the substrate,the prepared Ni7/Cu3-GO catalyst was found to be comparable to the noble metal catalysts reported in the references,as shown in Table 6.Compared with the noble metal catalyst,the catalyst dosage was about ten times that of the noble metal catalyst to achieve a similar catalytic effect(Table 6,entries 1-4).However,the noble metals Pd and Au were thousands of times more expensive than Cu and Ni.Compared with other reported non-precious bimetal catalysts,this Ni7/Cu3-GO catalyst exhibited pretty good catalytic performance even in the absence of alkali additives,accompanied by low consumption of hydrogen source (Table 6,entries 1,5,6).

4.Conclusions

In this study,Ni/Cu NPs with different Cu/Ni ratios were successfully synthesized by successive thermal decomposition of Ni(OAc)2·4H2O-OAM complex and Cuf-OAM complex with tunable molar ratio.OAM acted as the complexing agent and stabilizer,while liquid paraffin was used as solvent.The particle size of Ni/Cu NPs with different Cu/Ni ratios ranged from 40 to 45 nm,and the crystal structure of Ni/Cu NPs were close to the inner core fcc Ni.Ni/Cu-GO catalysts were prepared by anchoring Ni/Cu NPs onto graphene oxide (GO).Furthermore,the CTH reaction of nitroaromatics was investigated when ammonia borane was used as the hydrogen source.Cu dominated the main catalytic role in the reaction,while Ni played a synergistic role of catalysis and provided magnetic properties.The Ni7/Cu3-GO catalyst with the saturation magnetization of 11.3 emu·g-1was the best magnetic catalyst,showing 2-methyl-5-nitrophenol conversion and 5-amino-2-methylphenol yield of 99% and 96%,respectively.Moreover,the catalyst showed good reusability with the 5-amino-2-methylphenol yield of above 90% after six cycles of reaction.In addition,dehalogenation did not occur for the CTH of halogenated nitroaromatics.Based on the above results,this work reported a novel synthesis method of Ni/Cu NPs and Ni/Cu NPs anchored on GO,and provided a new and effective route for non-noble metalcatalyzed transfer hydrogenation of nitroaromatics with the strongly reductive hydrogen sources.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The work was supported by the National Natural Science Foundation of China (Grant No.21776161).